radiation effects on fiber opticsradiation effects on fiber optics 1. introduction 1.1 background...
TRANSCRIPT
AFCRl-TR-7 5-0190 PHYSICAL SCIENCES RESEARCH PAPERS, NO. 627
Radiation Effects on Fiber Optics
JAMES A. WALL JOHN F. BRYANT
2 April 1975
Approved for public: release; distribution unlimited.
Thi s research was supparted by the Air Force In-House Laboratory Independent Research Fund.
SOLID STATE SCIENCES LABORATORY PROJECT IliR
AIR FORCE CAMBRIDGE RESEARCH LABORATORIES HANSCOM AFB, MASSACHUSETTS 01731
AIR FORCE SYSTEMS COMMAND, USAF
8 .: \. ~
- - - . '.~ (,~UO - i ,~)-,
\,', \. -j
Unclassified SECURITY CLASSI FICATION OF THIS PAGE (When Data Entered)
REPORT DOCUMENTATION PAGE READ INSTRUCTIONS BEFORE COMPLETING FORM
1. REPORT NUMBER r GOVT ACCESSION NO. 3. RECIPIENT'S CATALOG NUMBER
AFCRL-TR-75-0190 4. TITLE (and Subtitle) 5. TYPE OF REPORT II PERIOD COVERED
RADIATION EFFECTS ON FIBER OPTICS Scientific. Final.
6. PERFORMING ORG. REPORT NUMBER
PSRP No. 627 7. AUTHOR(s) 8. CONTRACT OR GRANT NUMBER(s)
James A. Wall John F. Bryant
9. PERFORMING ORGANIZATION NAME AND ADDRESS 10. PROGRAM ELEMENT, PROJECT, TASK
Air Force Cambridge Res earch Laboratories (LQR) AREA 1\ WORK UNIT NUMBERS
HanscomAFB 61101F, ILIR3E
Massachusetts 01731 Related IHWU: ILIR3EOl
II. CONTROLLING OFFICE NAME AND ADDRESS 12. REPORT DATE
Air Force Cambridge Research Laboratories (LQR) 2 April 1975 Hanscom AFB 13. NUMBER OF PAGES
Massachusetts 01731 M 14. MONITORING AGENCY NAME & ADDRESS(iI dillerent from Controlling Olliee) IS. SECURITY CLASS. (01 thIs report)
Unclassified
158. DECL ASSI FI CA TI ON/ DOWN GRADI N G SCHEDULE
16. DISTRIBUTION STATEMENT (of this Report)
;Approved for public release; distribution unlimited.
". "''' ;""""" ""''''f ,., ~:."O<, '"""'7'Y~;'U ".0 .. ' - ~'"'" I, --_1.1 (9). '
,;I , 18. SUPPLEMENTARY NOTES
This work was a coo})erative effort between AFCRL and NELC. The entire effort was sUP1krted by AFCRL Laboratorlx Directorls funds. The NELC por-tion of the wor was performed under Con ract MIPR FY71217300017. A ~ortio~ of the work berformed by AFCRL under ARPA Order No. 2490 3DIO is a so ' included in t is report.
19, KEY WORDS (Continue on reverse side If necessary and Identify by block number)
Fiber optics Radiation effects Transients
20. ABSTRACT (Continue on reverse Bide II necessary and Identify by block number)
The replacement of conventional electrical cables with fiber optics tion transmission systems has the following advantages:
in informa-
1. No electromagnetic radifltion from, or pickup by, cables. 2. Extremely wide bandwidth. 3. Electrical isolation between transmitter and receiver. 4. Smaller size and weight by approximately an order of magnitude.
DD FORM I JAN 73 1473 EDITION OF I NOV 65 IS OBSOLETE
U nclas sHied SECURITY CLASSIFICATION OF THIS PAGE (When Date Entered)
TInclassified SECURITY CLASSIFICATION OF THIS PAGE(When Det. Entered)
These and other advantages are leading to widespread use of fiber optics in military command, communications, and control systems. One problem in the application of fiber optics to military systems is that nuclear (and space) radiation is known to produce color centers in optical materials, caus ing a reduction of light transmission over a spectrum of wavelengths. Also, energetic radiation can generate light within optical materials. These and other radiation effects could cause permanent or temporary disruption of a fiber optics transmission system.
Samples representative of most of the optical fibers presently available in lengths of 10 m or more were tested for their responses to energetic radiation. Glass fibers doped with three different levels of cesium were also prepared and tested. Permanent and transient x-radiation effects tests were performed using the AFCRL linear accelerator and flash x-ray machine. Neutron effects tests were performed using the fast burst reactor at White Sands, New Mexico.
All of the fibers tested showed decreases in transmission when exposed to radiation. Commercial grade glass fibers were found to be the most radiation sensitive. They became virtually,unusuable in lengths greater than 10 m after exposure to only a few hundred rads of x-rays. Considering all of the tests performed, germanium doped fused silica fibers proved to be the least radiation sensitive of the fibers tested. Although their transmission decreased more rapidly with increasing x-ray dose, the plastic fibers showed more rapid recovery from both long-term and transient radiation effects than the germanium doped fused silica fibers. The radiation sensitivities of the ces ium doped fibers were intermediate between the commercial grade glass and plastic fibers, with their radiation hardness increasing with increasing cesium doping.
All of the fibers emitted fluorescent light pulses when exposed to intense x-ray pulses. The fluorescent intensity emitted at the ends of the fibers, per unit length of irradiated fiber, was greatest' for the commercial grade glass fibers and least for the fused silica fibers. The duration of the fluorescence following an x-ray pulse ranged from less than a microsecond for the plastic and fused silica fibers to several microseconds for the commercial grade and cesium doped glass fibers.
Unclassified SECURITY CLASSIFICATION OF THIS PAGE(When D.t. Entered)
1. INTRODUCTION
1. 1 Background 1. 2 Program Scope and Objectives
2. EXPERIMENTAL PROCEDURES
2.1 Technical Approach 2. 2 Experlmental Setup 2. 3 Operational Procedures
2.3.1 Permanent X-Ray Effects 2.3.1.1 Irradiation of Fibers 2.3.1.2 Pre- and Post-Irradiation Spectral
Transmission 2.3.2 Neutron Effects 2.3.3 Transient X-Ray Effects
3. RESULTS
3.1 Data Treatment 3.2 Permanent X-Ray Effects
3.2.1 Commercial Grade Glass Fibers 3.2.2 Low-Loss Fused Silica Fibers 3.2.3 Plastic Fibers 3.2.4 Cerium Doped Fibers 3.2.5 Relative Sensitivities of Fibers to X-Rays
3.3 Neutron Effects 3.3.1 Commercial Grade Glass Fibers 3.3.2 Low-Loss Fused Silica Fibers 3.3.3 Plastic Fibers 3.3.4 Cerium Doped Fibers 3.3.5 Relative Sensitivities of Fibers to Neutrons
3.4 Transient X-Ray Effects 3.4.1 Transient Change in Transmission 3.4.2 Fluorescent Pulse
3.5 Summary of Results
3
Contents
7
7 8
9
9 10 12 12 12
14 14 15
16
16 18 18 22 25 30 33 34 34 36 37 38 40 42 42 45 49
Contents
4. CONCLUSIONS AND RECOMMENDATIONS 53
Illustrations
1. Schematic of Experimental Arrangement
2. Change in Transmission Versus Dose, Bendix K.2K
3. Change in Transmission Versus Dose, Rank Hi-Tran
4. Change in Transmission Versus Dose, Schott LK
5. Attenuation Versus Wavelength, Bendix K2K
6. Attenuation Versus Wavelength, Rank Hi-Tran
7. Attenuation Versus Wavelength, Schott LK
8. Change in Transmission Versus Dose, Corning Low-Loss "A"
9. Change in Transmission Versus Dose, Corning Low-Loss "B"
10. Change in Transmission Versus Dose, Corning 1-279
11. Attenuation Versus Wavelength, Corning Low Loss "A"
12. Attenuation Versus Wavelength, Corning Low Loss "B"
13. Change in Transmission Versus Dose, Dupont Crofon® 1110
14. Change in Transmission Versus Dose, International Fiber Optics 1032
15. Relative Induced Transmission Loss Versus Time After Irradiation, International Fiber Optics 1032
16. Attenuation Versus Wavelength, Dupont Crofon ® 1110
17. Attenuation Versus Wavelength, International Fiber Optics 1032
18. Change in Transmission Versus Dose, Cerium Doped Glass Fibers
19. Attenuation Versus Wavelength, 0.1 percent Cerium Doped Glass Fiber
20. Attenuation Venous Wavfllength, 0.2 percent Cerium Doped Glass Fiber
21. Attenuation Versus Wavelength, 0.3 percent Cerium Doped Glass Fiber
22. Change in Transmission Versus Neutron Fluence, Bendix K2K
23. Change in Transmission Versus Neutron Fluence, Corning 5011
24. Change in Transmission Versus Neutron Fluence, Corning Low-Loss "B"
25. Change in Transmission Versus Neutron Fluence, Dupont Crofon@ 1110
4
10
19
19
20
20
21
21
22
23
24
25
26
26
27
28
29
30
31
31
32
32
34
35
37
38
Illustrations
26. Change in Transmission Versus Neutron Fluence, 0.1 percent Cerium Doped Glass Fiber 39
27. Change in Transmission Versus Neutron Fluence, 0.3 percent Cerium Doped Glass Fiber 40
28. Signal Transmitted by 4.5 m Length of Dupont Crofon ® Plastic Fiber Irradiated With 1200 rad, 20 nsec X-ray Pulse 42
29. Signal Produced by Fluorescent Pulse Generated in 9.8 m Length of Corning 1-279 Single Fiber Irradiated With 1000 rad, 20 nsec X-ray Pulse 45
Tables
1. Comparison of Sensitivities of Fibers to X-Radiation 33
2. Comparison of Sensitivities of Fibers to Neutrons 41
3. Characteristic Constants of Eq. (6) for Transient Radiation Induced Change in Transmission of Fibers 43
4. Fluorescent Pulse Characteristics 48
5. X-Ray Dose Required to Reduce the Transmission of a 30 m Length of Optical Fiber to 50 percent of its Original Value 50
6. Neutron Fluence Required to Reduce the Transmission of a 30 m Length of Optical Fiber to 50 percent of its Original Value 50
7. Time Required for the Transmission of a 30 m Length of Fiber to Recover to 50 percent of its Original Value Following a 1000 Rad X-Ray Pulse 51
8. Fluorescent Radiant Intensity at the End of a 30 m Length of Fiber Exposed to an X-Ray Pulse 52
5
Radiation Effects on Fiber Optics
1. INTRODUCTION
1.1 Background
Interest in replacing conventional electrical cables used in information trans
mission and control systems with fiber optic light guides has grown considerably
in recent years. Transmission systems using fiber optics are similar to those
using electrical cables except that the output of the transmitter is used to modulate
the intensity of a light source. The source is coupled to one end of an optical
fiber which replaces the electrical cable. The other end of the fiber is coupled to
a photosensitive device (detector). The detector responds to the light transmitted
by the fiber, producing an electrical signal corresponding to the output of the
modulated light source. This signal can then be processed as in a system using
conventional cables. Some of the advantages of the use of fiber optics in trans
mission systems are:
(1) No electromagnetic radiation from, or pickUp by, cables
(2) Extremely wide bandwidth
(3) Electrical isolation between transmitter and receiver
(4) Smaller size and weight by approximately an order of magnitude
These and other advantages are leading to widespread use of fiber optics in
military command, communications, and control systems.
(Received for publication 31 March 1975)
7
One problem in the application of fiber optics to military systems is that
nuclear (and space) radiation is known to produce color centers in optical materials,
causing a reduction of light transmission over a spectrum of wavelengths. Also,
energetic radiation can generate light within optical materials. Thes e, and other
radiation effects could cause permanent or temporary disruption of a fiber optics
transmission system. Conventional optical systems contain no more than a few
cm of optical material in the light path so radiation effects may be negligible at
moderate doses and dose rates. However, systems using fiber optics may have
several m, some as much as a km or more, of optical material between the light
source and detector. In such systems radiation can produce severe changes in the
operation of a system even at relatively low doses and dose rates.
1.2 Program Scope and Objectives
There has been extensive work done in the past on the effects of nuclear radia
tion on optical materials. 1, 2, 3 However, when this project began there was essen
tially no information on how radiation affected the performance of optical fibers.
The investigation of radiation effects on bulk optical materials that might be used
to produce radiation hard optical fibers was considered premature until it was
determined whether any effects might occur in the fibers that could not be predicted
from prior work. The program was therefore restricted to tests of available
optical fibers.
There are many systems, ranging from space vehicles to land based instal
lations, that may utilize fiber optics. Each system mus t meet radiation tolerance
levels below which it must continue to operate for a specified minimum period of
time. The lowest tolerance level is for manned systems such as aircraft, land
based communications systems, and naval vessels. These are also the systems
in which fiber optics are most likely to be used in the near future. The radiation
effects tests were therefore directed primarily toward this tolerance level, which
was assumed to be 2000 rads.
Within the above limits, the objectives of this program were:
(1) Evaluate available optical fibers, both commercial and experimental, for
their applicability to manned and other systems that may be exposed to energetic
radiation.
1. Monk, G. S. (1950) The Coloration of Optical Materials by High Energy Radiations, Argonne National Laboratory - 4536.
2. Bishay, Adli (1970) Radiation induced color centers in multicomponent glasses J. Noncrystalline solids ::":54-114.
3. Lell, E., Kreidl, N. J., and Hensler, J. R. (1966) Radiation Effects in Quartz, Silica, and Glasses, Progress in Ceramic Science Vol. 4, Pergamon Press, Oxford.
8
(2) Determine the requirements for a development program to produce radia
tion hard optical fibers if needed.
(3) Form a basis to establish radiation effects test procedures for optical
fibers.
2. EXPERIMENTAL PROCEDURES
2.1 Technical Approach
The radiation induced transmission losses in optical materials is known to be
a nonlinear function of absorbed dose. Because of this nonlinearity the results of
radiation effects tests performed on short lengths (1-2 m or less) of optical fibers
cannot be reliably extrapolated and applied to systems employing fiber lengths of
30 m or more. It was therefore decided to perform tests on fiber lengths in the
range of 10 m. This choice also restricted the tests to those optical fibers being
produced in such minimum lengths.
A "live" test procedure was adopted for the permanent damage tests. During
these tests the transmission of the fibers was observed continuously during the
irradiations. This made possible the observation of any effects, such as the rapid
annealing of radiation induced damage, that might occur within time periods of
seconds or minutes. Such effects would ordinarily go unnoticed in simple pre
and post-irradiation tests.
The permanent damage tests were performed over only a narrow band of wave
lengths corresponding to the emission of a light emitting diode. In order to deter
mine if there were any wavelength regions where radiation induced transmission
losses would be minimized, spectral transmission measurements were made on
the fibers before and after the irradiations. These spectral measurements could
also be used to compare the losses induced in the fibers with published radiation
damage data on bulk optical materials to determine if the fiber drawing processes
introduced any effects peculiar only to the optical fibers.
Transient radiation effects tests; that is, the responses of the fibers to single
high intensity bursts of radiation, were performed with and without light being
transmitted through the fibers. This made possible observation of radiation in
duced fluorescence as well as short term losses of transmission in the fibers.
Although attempts were made to determine the spectral distributions of the
fluorescence and transmission losses by use of a scanning optical spectrometer,
electronic malfunctions that could not be corrected before completion of the pro
ject made these measurements impossible.
9
2.2 Experimental Setup
Figure 1 is a schematic of the experimental arrangement used for the radia
tion effects tests of the optical fibers. The lengths of fiber were formed into
circular coils with outer diameters of approximately 10 cm to insure uniform ir- .
radiation over mos t of the fiber length. One end of the fiber was coupled to a
light source and the other end to a photodetector. The source and detector were
located in a radiation shielding enclosure that was positioned as far away as prac
tical from the radiation source. The output of the photodetector was connected to
an oscilloscope located in the control area. This permitted continuous monitoring
of the magnitude of the optical signal transmitted by the fiber during irradiation.
All of the fibers tested were ·jacketed multifiber bundles (wi th the exception
of a Corning 1-279 experimental fiber which was a single fiber). Brass ferules
were cemented to the ends of the bundles and the fiber ends ground and polished.
Optical coupling to the light source and photodetector was by contact only, no
lenses or coupling compounds being used.
For the permanent damage tests the light source used was a pulse modulated
light emitting diode (LED) as indicated in Figure 1. The LED was driven by a
switching circuit (a Texas Instruments SN7404 integrated circuit) to obtain light
pulses of uniform amplitude. The switching circuit was triggered by a pulser,
located in the control area, that could be adjusted to set the repetition rate and
RADIATION --~
SHIELD
IRRADIATION AREA
CONTROL AREA
Figure 1. Schematic of Experimental Arrangement
10
width of the light pulses. An infrared (905 nm) emitting LED (Monsanto ME5) was
used during tests of the glass optical fibers. This was replaced by a visible
(660 nm) emitting LED for testing plastic fibers which have intense absorption
bands in the 900 nm region.
For the transient radiation effects tests the pulse modulated LED was replaced
by a high-intensity monochromator (Bausch and Lomb 33-86 series). This pro
vided a constant light source which eliminated the need for pulse synchronization
with the single high intensity radiation pulses used in these tests. For glass fibers
the monochromator was set for 905 nm and for the plastic fibers 660 nm (both with
nominal 10 nm bandwidths) to correspond to the LED emissions used in the per
manent damage tests.
The photodetector used was a Texas Instruments TIXL79 avalanche photodiode
module. This module has a built-in video amplifier with a bandwidth of 40 MHz.
The module responsivity is 2 X 105 volts /watt at 900 nm.
The radiation sources used in the tests were the AFCRL linear accelerator
and flash x-ray machine and the fast burst reactor at White Sands Missile Range,
New Mexico. The linear accelerator was used for the permanent damage x-ray
effects tests. 10 MeV electrons from the accelerator incident on a tungsten target
produced 4. 5 J.lsec wide pulses of x-rays. The pulse repetition .rate could be
varied from manually triggered single pulses to internally generated pulse trains
ranging from one to 180 pulses per second. In the position relative to the x-ray
target where most of the fibers were irradiated the fibers received about 2 rads
(SO per pulse.
The flash x-ray machine produced bursts of x-rays with a pulse width of
20 nsec. The dose received by the fibers was determined by the distance between
the fiber coil and the (internal) machine x-ray target. The doses used in the testf
ranged from 20 to 1200 rads per pulse.
The fas t burs t reactor produced burs ts of mixed neutrons and gammas. The
bursts were gaussian in shape with a nominal half width of 50 J.lsec. The neutron
fluence received by the fibers was determined by the settings of the reactor con
trol rods, which could be changed from pulse to pulse, and the position of the
fiber coil relative to the reactor. For thes e tes ts the fibers were located at a dis
tance (100 cm) from the reactor where they would receive neutron fluences rang
ing from 1011 to 1012 neutrons/cm2. The dose received by the fibers due to the
gamma radiation from the reactor was typically 2 X 10-10
rad/neutron/cm2
•
The radiation shield indicated in Figure 1 was required to prevent the radia
tion from reaching the light source and photodetector where it could produce false
signals due to transient effects or gradual degradation of the performance of the
equipment during prolonged use. For irradiations performed with the linear ac
celerator the shield wall was a 5 cm thickness of lead; for the flash x-ray machine
11
it was 10 cm of lead; and for the fast burst reactor it was a composite of about
30 cm of polyethylene, 15 cm of lead, and O. 3 cm of boral.
Thermoluminescent dosimeters (TLD's, CaF imbedded in teflon) were used
for dose measurements associated with the linear accelerator and flash x-ray
machine. Also, a PIN diode was used for dose-rate monitoring of the linear ac
celerator and a phosphor-photodiode combination was used to monitor the dose
delivered by the flash x-ray machine. Dosimetry at the White Sands fast burst
reactor was provided by that facility and employed sulfur pellets for neutron
dosimetry and thermoluminescent LiF powder for gamma dosimetry.
Some of the pre- and post-irradiation spectral transmission data on the fibers
were taken using" available" spectroscopic equipment pending receipt of a per
manent setup. However, most of the spectral data shown in this report were taken
using a Bausch and Lomb series 33-86 grating monochromator, silicon photodiode,
and PAR lock-in amplifier. An American Time products type 40 tuning fork
chopper, operating at 100 Hz, was located between the tungsten-iodide light source
and the entrance slit to the monochromator. One end of the fiber optic bundle
was positioned at the exit slit of the monochromator and the other end was coupled
to the silicon photodiode. The output of the photodiode was connected to the input
of the lock-in amplifier, which was synchronized with the output of the tuning fork
drive circuit. Although the entrance and exit slits were set to obtain a 10 nm
bandwidth, the actual bandwidth was determined to some degree by the effective
apertures of the fibers. However, since only relative spectral transmission data
was desired rather than absolute measurements of transmission, this arrangement
was adequate for these tests.
2.3 Operational Procedures
2.3.1 PERMANENT X-RAY EFFECTS
2. 3. 1. 1 Irradiation of Fibers
Using the experimental setup described above and shown in Figure 1, the coil
of optical fibers was positioned in front of the linear accelerator x-ray target.
TLD's were taped to the coil along the vertical and horizontal axes on both the
front (facing the target) and back of the coil to obtain a measure of .the average dos e
that would be accumulated by the fiber. The LED was driven with a one J.lsec wide
rectangular pulse with a repetition rate of approximately 200 kHz. Before the
irradiation began the signal transmitted by the fiber was observed on the oscil
loscope and photographed.
At the beginning of the irradiation the accelerator was manually triggered to
obtain single x-ray pulses. During the pulsing the amplitude of the signal trans
mitted by the fiber was continuously observed on the oscilloscope. When any
12
significant change occurred in signal amplitude the irradiation was stopped just
long enough to photograph the oscilloscope trace. When a fiber did not show a sig
nificant change in transmitted signal during the manual pulsing, the accelerator
was switched to the internal trigger mode and the pulse rate was slowly increased
until the signal began to change at an acceptable rate. As with the manual trig
gering, the irradiation was stopped to photograph significant changes in signal am
plitude. The process of stopping the irradiation to photograph the signal required
only a fraction of a second-the time required to switch off the accelerator trigger,
trip the camera shutter, and switch on the trigger.
During the irradiations the number of x-ray puIs es produced by the accelerator
was accumulated on a pulse counter driven by the accelerator trigger circuit. Each
time the signal on the oscilloscope was photographed the number of pulses ac
cumulated up to that time was recorded. This information was later used to com
pute the dose (received by the fiber) that produced the recorded change in signal
amplitude. This was done by multiplying the number of pulses generated by the
accelerator up to the time the signal was photographed by the dose received by the
fiber per x-ray pulse. The dose per pulse was obtained by dividing the total dose
accumulated by the fiber during the irradiation, as determined from the TLD's
taped to the fiber, by the total number of pulses generated during the irradiation.
The PIN diode mentioned in paragraph 2.2 was also used to check the dose per
pulse by putting it in place of the fiber and operating the accelerator under the
same conditions as during the irradiation.
The irradiations were occasionally interrupted long enough to look for any
short-term annealing; that is, recovery of transmission loss, that might occur in
a fiber. If annealing was observed the signal trace was photographed at intervals
as changes in amplitude occurred. The time between photographs was recorded to
obtain data on the annealing process as a function of time.
Also during the irradiations the LED trigger puIs e was occas ionally turned
off to look for fluorescence generated in the fibers by the x-ray pulses. No such
fluorescence was observed, however,
The irradiation of a fiber was usually continued until the signal transmitted
had dropped in amplitude to between 25 and 50 percent of its original value. The
fibers were not usually irradiated to the point where the signal to noise ratio made
measurement of the signal amplitude impractical in order to permit measurement
of the spectral transmission of the fiber without cutting it to shorter lengths. Cut
ting the fibers after irradiation in order to make the spectral measurements was
not considered desirable because this would mean putting on new end-fittings and
repolishing the ends. This, along with possible breakage of some of the fibers
in the bundle, could affect the results of the measurements. Only one fiber, the
Corning low-loss type A, was irradiated to the point where it had to be cut in order
13
to make the spectral measurements. When it was desired to determine the induced
losses of other fibers as a function of dose at doses greater than required to pro
duce a 50 to 75 percent drop in signal amplitude, the fibers were reirradiated some
time after the spectral data was obtained following the first irradiation.
2.3.1.2 Pre- and Post Irradiation Spectral Transmission
Pre- and post-irradiation spectral transmission distributions were measured,
with some exceptions as noted below, on the same coils of fibers that were used
in the permanent x-ray effects tests. Using the equipment described in paragraph - -
2.2 monochromator-photodiode "source" spectral distributions were measured by
butting together two blank ferules of the type used to terminate the fiber bundles,
connecting them between the monochromator and photodiode, and measuring the
output of the photodiode as a function of wavelength at 10 nm intervals. The
ferules were then replaced with the coiled optical fiber and the process repeated.
Dividing the results obtained with the fiber in place by those obtained with the
ferules in place gave the relative spectral transmission of the fiber. This pro
cedure was not intended to give an absolute measure of fiber spectral transmission
but only to indicate the changes that occurred due to the irradiations.
The exceptions to this procedure were the Corning low-loss type A as men
tioned in paragraph 2.3.1. 1, the Corning low-loss type B, and the plastic fibers.
The type A transmission spectra were measured by comparing the transmission of
a 2 m length of fiber with the trans miss ion of a O. 5 m length, the shorter lengths
being cut from the coil of fiber before and after irradiation. The post irradiation
transmission spectrum of the type B fiber was measured on the full irradiated coil
as described above, but the spectrum for the un-irradiated fiber was taken using
2 and 0.5 m lengths taken from the same lot of fiber (this was necessitated by a
tight schedule between irradiation sources and a limited supply of the fiber). The
plastic fiber transmission spectra were also measured using the shorter lengths
of fiber, but the same pieces of fiber were used for both the pre- and post
irradiation measurements. This was done by combining the short lengths of fiber,
on which spectral measurements had been made, with the longer coils of fiber
during the irradiations. The reason was that these fibers have very intense
absorption bands at wavelengths greater than 800 nm. The transmission of the
fibers at these wavelengths was too low for measurements to be made on longer
lengths of fiber so the shorter lengths were used in order to determine the effects
of radiation on the absorption bands.
2.3.2 NEUTRON EFFECTS
The experimental arrangement used at the fast-burst reactor for the neutron
effects tests was similar to that used for the permanent x-ray effects tests except
as noted in paragraph 2.2. Two oscilloscopes with their inputs connected in
14
parallel were used to monitor the output of the photodetector. The horizontal
sweep trigger of one of the oscilloscopes was delayed approximately 100 fJ,sec with
respect to the other (which had no delay) to enable observation of the fiber trans
mission immediately following a reactor burst. Prior to a burst the oscilloscopes
were set for continuous sweep operation and the trace of the signal transmitted by
the fiber was photographed. The oscilloscope sweeps were then set to be triggered
by a pulse synchronized with the reactor trigger. The oscilloscope carrlera shutters
were manually opened about 1 sec before the reactor was to fire and clos~d follow
ing the burst. This gave photographs of the signal transmitted by the fiber during
and immediately following the burst. However,. it was found that the electrical
noise level during the bursts was very high; so photography of the signal during
the bursts (the oscilloscope trace without any time delay) was abandoned although
visual observation of that signal was maintained.
Each fiber was exposed tr. two to four reactor bursts. The sulfur pellets and
TLD's used for neutron fluence and gamma dose measurements were left attached
to the coil of optical fiber throughout the irradiations so that they measured the
accumulated fluence and dose delivered to fiber. The neutron fluence produced by
the reactor is directly proportional to the temperature rise of the reactor core
during a burst. Photographs, taken by the reactor operating personnel, showing
oscilloscope traces of the temperature rise and burst pulse shape were obtained
for each reactor burst. Using this information along with the dosimetry data,
the neutron fluence at the fiber position was computed for each burst. The gamma
dose to the fiber for each burst was similarly obtained since the gamma-neutron
ratio is a constant for a fixed position relative to the reactor centerline.
Following each reactor burst the signal transmitted by the fiber was checked
at various time intervals to see if any annealing of induced transmission losses
occurred. The period over which these observations could be made was usually
determined by the minimum required waiting time between reactor burs ts. This
ranged from 30 min for the lower neutron fluence bursts to 1 hr for the maximum
fluence bursts.
2.3.3 TRANSIENT X-RAY EFFECTS
The pulse modulated LED used in the permanent x-ray and neutron effects
tests was replaced by a monochromator as mentioned in paragraph 2.2. Although
the tungsten-iodide lamp used with the monochromator was powered from a 60 Hz
ac line, it provided an essentially constant light source. The coil of fiber to be
tested was coupled between the exit slit of the monochromator and the photo
detector. The output of the photodetector was connected to the dc input of a wide
bandwidth oscilloscope (Tektronix 7904). With the oscilloscope operating on con
tinuous sweep,' the displacement of the trace from the zero input signal position
15
indicated the optical signal level transmitted by the fiber. Although the mono
chromator entrance and exit slits were set for a nominal 10 nm bandwidth, if a par
ticular fiber did not have enough transmission to provide a signal level adequate
for good sensitivity in the tests the entrance slit was widened to provide sufficient
signal (the exit slit was not adjustable). Also, if the signal was so large that it
could saturate the photodetector. the entrance slit width was reduced to bring the
signal within the linear dynamic range of the detector.
As in the permanent x-ray effects tests, TLD's were taped to the fiber coil
to measure accumulated dose. The coil was usually positioned relative to the x-ray
target so that it would receive approximately 100 rads from a single x-ray pulse.
The oscilloscope was set for stngle sweep operation that could be triggered by a
synchronizing pulse from the flash x-ray machine. Before the irradiation, the
oscilloscope sweep was triggered by the flash x-ray machine with no x-ray output
in order to obtain a photograph of the signal level transmitted by the fiber under
actual experimental conditions. The machine was then fired to obtain a photograph
of the transient absorption induced by an x-ray pulse. Fibers that did not show a
measurable response to the x-ray burst were moved closer to the x-ray target and
irradiated with a more intense pulse. The same procedure was followed to measure
fluorescence induced in the fibers by x-ray pulses except that an opaque shield
was placed between the light source and entrance slit of the monochromator to cut
off light input to the fiber.
3. RESULTS
3.1 Data Treatment
All results on radiation induced changes in fiber transmission are reported in
units of dB/km. The change in transmission was computed from the relation:
.6T (1)
where
.6T = change in transmission in dB/km
S = amplitude of signal transmitted by fiber before irradiation o S = amplitude of signal after irradiation
L = total length of fiber tested
1. = unirradiated length of fiber used as leads between light source and
photodetector.
16
Pre- and post-irradiation spectral measurements are shown as plots of
attenuation in dB/km versus wavelength. The attenuation was computed from an
equation identical to (1) by redefining the terms as follows:
So = signal measured at wavelength 1\ with shorter length of fiber in
spectrometer
S = signal measured at wavelength 1\ with longer length of fiber in
spectrometer
£ = shortest length of fiber in km
L = longes t length of fiber in ~m
For spectral measurements on the full fiber coils used in the tests S becomes the o source detector function of the spectrometer and £ = O.
As previously stated in paragraph 2.2, the spectral transmission measure
ments were intended to yield only relative attenuation data. In order to show the
relative change between the pre- and post-irradiation spectra in relation to the
change in transmission observed during irradiation, the post-irradiation spectra
were adjusted relative to the pre-irradiation spectra. This was done by taking
the difference (in dB/km) between the two spectra at the light-source wavelength
used during the irradiation. This was compared with the change in transmission
observed at the end of the irradiation. The entire post-irradiation spectrum was
then adjusted by adding or subtracting the appropriate number of dB/km to obtain
agreement between the spectral results and the irradiation results at the appro
priate wavelength. This correction was usually less than 10 percent of the attenua
tion measured in the post-irradiation spectrum at the wavelength in question.
The optical fibers tested can be classified into four types as follows:
(1) Commercial Grade Glass Fibers. These are general purpose fibers
available as stock items from manufacturers. They are drawn from "standard"
optical glasses such as lead silicate and have intrinsic attenuations of the order of
1000 dB/km.
(2) Low-loss Fused Silica Fibers. These are drawn from pure fused silica
with a dopant added to the core to increas e its refractive index. Intrins ic attenua
tion is less than 100 dB/km.
(3) Plastic fibers. These are very inexpensive fibers with cores of either
polymethylmethacrylate or polystyrene and a clad of lower refractive index poly
mer. They have very high attenuation at wavelengths greater than 800 nm. Their
attenuation in the visible is about 1200 dB/km.
(4) Cerium-doped fibers. These are glass fibers made by adding a small
amount of cerium oxide to the glass melt before they are drawn to increase their
res istance to radiation induced trans miss ion losses. (Cerium doped glass has
been used for several years to produce hot-cell windows and optical components
17
for use in high-level radiation areas). The cerium doped fibers tested were
specially made for this project and had initial attenuations of approximately
1200 dB/km.
The radiation effects observed were distinctly different for each type of fiber.
The results of each radiation test are therefore reported according to fiber type.
3.2 Pel'manent X-Ray Effects
3.2.1 COMMERCIAL GRADE GLASS FIBERS
Three commercial grade glass fibers we're tested. These were the Bendix
K2K, ;" Rank Hi-Tran, and Schott LK fibers. The results of the radiation tests are
shown in Figures 2, 3, and 4 where the observed change in fiber transmission is
plotted against the dose received by the fiber. The fibers were initially irradiated
until their trans miss ions dropped to about 50 percent of the pre-irradiation levels.
Forty days later the fibers were recoupled to the same LED and photodetector and
irradiated again until the transmitted signal was too s mall for accurate measure
ments to be made. The time interval between irradiations is indicated on the plots.
As can be seen in the figures, the Bendix K2K fiber had regained some of its trans
mission loss during the time interval between irradiations. The Rank Hi-Tran
'showed no significant change in transmission during the time interval while the
Schott LK data indicate a decrease in transmission. This latter effect, however,
could be and probably is due to a slight difference in the coupling of the fiber to the
LED and photo detector between the two irradiations.
Least-square fits were made to these data and are shown as the straight lines
through the data points on the plots. The change in transmission is a closely linear
function of dose over the range of exposures used in these tests. (Since change in
transmission is expressed here in dB/km, this means that the amplitude of the
transmitted signal changes exponentially with dose). The slopes of these linep,
indicated on the plots, can be used as a relative measure of the radiation hardness
of the fibers. The slopes of the lines through the Bendix and Rank fiber data are
very nearly the same and are significantly greater than the slope of the line through
the Schott fiber data. Although the Schott fiber is less sensitive to radiation than
the other two, it is clear from the data that none of these fibers would remain
useable (except in much shorter lengths than tested) if exposed to the 2000 rad
dose tolerance limit assumed for manned systems.
Figures 5, 6, and 7 show the pre- and post-irradiation spectral transmissions
of the Bendix, Rank, and Schott fibers respectively. The pre-irradiation spectra
'~The Bendix optical fiber production facilities had been taken over by the Galileo Electro Optics, Inc., when these tests were performed. However, the Bendix name is retained in this report because the fiber tested was so labeled.
18
-400
E-350 ~ CO
~-300 z o ~-250 ::!: en ~-200 0:: I-
~-150
w (!)
~-100 :r: u
-50
SLOPE = 2.53 dB/km-RAD ----+'
SLOPE = 2.49 dB/km-RAD •
\ FIBER BUNDLE LENGTH:29 METERS RADIATION: 10 MeV X-RAYS
• SOURCE WAVELENGTH: 905 nm
? / . A 1 sl IRRADIATION
40 DAYS ELAPSED TIME • 2nd IRRADIATION
O~ __ ~ __ ~ __ ~ __ ~ __ -L __ -L __ -L __ -L __ ~
o 20 40 60 80 100 120 140 160 180 DOSE (RADS)
Figure 3. Change in Transmission Versus Dose, Rank Hi-Tran
E .>< "-co '0
z 0 iii en ::!: en z <t 0:: I-
~ w (!)
z <t :r: u
-550
-500
-450
-400
-350
-300
-250
-200
-150
-100
-50
40
19
80
Figure 2. Change in Transmission Versus Dose, Bendix K2K
+-SLOPE =2.44 dB/km-RAD
<-40 DAYS ELAPSED TIME
FIBER BUNDLE LENGTH: 14 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH: 905nm
120 160 200 240 280 320
DOSE (RADS)
Figure 4. Change in Transmission Versus Dose, Schott LK
-550
-500
-450
E ~ -400 CD ."
z -350 o en en i -300 en z ~ -250 f-
z ;:; -200 (!)
z <! 5 -150
-100
+- SLOPE = 1.53 dB/km-RAD
1"'- 40 DAYS ELAPSED TIME
FIBER BUNDLE LENGTH:13 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH:905nm
A 1st IRRADIATION
• 2nd IRRADIATION +- SLOPE = 1.67 dBI km- RAD
3000
E -" ..... CD ~2000 z o ~ ::J Z W f- 1000 ~
+-- AFTER DOSE OF 46 RADS
1 BEFORE IRRADIATION
o L--6-0l0--~--~8~0~0---L--~10~0~0~J-~1~2~0~0~ WAVELENGTH (n m)
20
100 200 300 400 DOSE (RADS)
Figure 5. Attenuation Versus Wavelength, Bendix K2K
E ~
"m "0
3000
;2000 o
~ :::J Z w I-~ 1000
+--AFTER DOSE OF 92 RADS
LBEFORE IRRADIATION
600 1000 WAVELENGTH (nm)
Figure 7. Attenuation Versus Wavelength, Schott LK
1200
Figure 6. Attenuation Versus Wavelength, Rank Hi-Tran
3000 ~RADS
~~ ~ BEFORE IRRADIATIO~ z o
~ :::J Z W
2000
~ 1000 «
0L---6JO-0---L---8JO-0---L---10JO-0---L---12JO-0--~
WAVELENGTH (nm)
21
are all very similar, with absorption bands near 650 and 970 nm. This similarity
suggests that the core glass is of one type for all three fibers. Differences in
radiation response between the fibers such as the greater radiation hardness of
the Schott and the annealing characteristic of the Bendix are probably due to minor
differences in the elemental compositions of the glasses. The post-irradiation
spectra show a greater increase in attenuation at the shorter wavelengths than at
longer wavelengths. This is typical of the radiation induced attenuation in most
glasses and suggests that the results of radiation effects tests of bulk glaElses can
be applied to optical fibers drawn from them. This of course assumes that no im
purities are introduced during the drawing process.
3.2.2. LOW-LOSS FUSED SILICA FIBERS
Three types of low-loss fused silica fibers, all manufactured by Corning
Glass Works, were tested. Two of these were designated by the manufacturer as
low-loss type A and low-loss type B and were supplied as jacketed multiple-fiber
bundles. The third was an experimental prototype of the low-loss type B. This
was a single fiber labeled as 1-279. Although the compositions of these fibers are
not known in detail, it is known that the core of the type A fiber is doped with
titanium while the cores of the type Band 1-279 are doped with germanium.
Figure 8 shows the results of the irradiation of the low-loss Type A. No data
was taken below the 220 rad dose level, but for the higher doses the change in
trans miss ion appears to be a linear function of dose. (The irradiation was actual
ly continued to a dose of 1000 rads although this is not shown in Figure 8). A least
squares fit to the data above 220 rads gave a slope of 0.74 dB/km-rad as indicated
Figure 8. Change in Transmission Versus Dose, Corning Low-Loss "A"
E .>< "-III "0
z 2 en en :::;: en z « a:: I-
~ w (,!) z « J: u
-800
-700
-600
-500
-400
-300
-200
-100
0 /
0
22
/ /
"
/ /
/
SLOPE =0.74 dB/km-RAD----+
FIBER BUNDLE LENGTH: 13.4 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH:905nm
/"- SLOPE =0.95 dB/km-RAD
100 200 300 400 500 600 700 800 DOSE (RADS)
on the plot. This fiber is therefore less than half as sensitive to radiation as the
commercial glass fibers but still is not sufficiently radiation resistant to be used
in manned systems.
The test results for the low-loss type B are shown in Figure 9. The response
of this fiber to radiation is clearly different than either the commercial glass
fibers or the low-loss type A. It is much more resistant to radiation damage than
the others. The initial slope of the curve drawn through the data points is ap
proximately 3 X 10-3 dB/km-rad, about three orders of magnitude less than the
other fibers. Some annealing of the radiation induced transmission loss was also
found and is indicated in Figure 9 by the data points at 20 kilorad and 160 kilorad.
After the fiber had received a dose of 20 kilorad, the irradiation was stopped for
15 min and the induced transmission loss was observed to drop from 58 to 50 dB/km.
The irradiation was then continued until the fiber had received a total dose of
160 kilorad. At this point the signal transmitted by the fiber had dropped in
amplitude to about 50 percent of its original magnitude, corresponding to a change
in transmission of 228 dB/km.Seventeen hours later the signal amplitude had
recovered to 87 percent of its original value. This represents a transmission loss
of only 50 dB/km, a 178 dB/km decrease in loss since the irradiation was stopped.
The LED was turned off during the 17 hr waiting period. The observed annealing
was therefore intrinsic to the fiber and not induced or enhanced by the light source.
The nonlinearity of the change in transmission as a function of dost ..:;f the
low-loss type B also distinguishes it from the commercial glass and type A fibers.
(However, thos e fibers could not be irradiated to the dos e levels received by the
type B because of their rapid loss in transmission. If they had been, they
Figure 9. Change in Transmission Versus Dose, Corning Low-Loss "B"
E .>< "-
-280
!g-240
z ~-200 (/) (/)
i (/)-160 z <t a:: f--120 ~ w (!) -80 z <t J:
u -40 • 15 MINUTE WAIT
FIBER BUNDLE LENGTH:13 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH:905nm
O~ __ -L __ -L __ ~ ____ L-__ -L __ -L __ ~ ____ L-~
o 20 40 60 80 100 120 140 160 180 DOSE (KILORADS)
23
might also have shown non-linear responses.) As will be seen below from the
results obtained in the tests of plastic fibers, an apparent nonlinear response
could result from a combination of the rate at which the transmission loss an
nealed and the rate at which the dose was delivered to the fiber. During the irra
diation of the type B fiber the linear accelerator was operated at a puIs e repetition
rate of about 20 pulses per second with a dose of 2 rads per pulse at the fiber posi
tion. It is possible that during the 50 ms period between pulses some of the loss
induced by the individual x-ray pulses annealed out. For this to cause the observed
nonlinearity the annealing rate immediately following an x-ray pulse would have
to be much greater than the rate observed when the irradiation was stopped.
Figure 10 shows the results obtained for the I-279 fiber. This fiber was
tested especially for the Defense Advanced Research Projects Agency and the 4 results of those tests are reported elsewhere. The data is shown here for com-
parison with the type B low-loss fiber. The greater scatter in the data points ob
taineo with this fiber as compared with the others is probably due to the sensitivity
of the coupling of the single fiber to its position relative to the light source and
photodetector. Minor vibrations and air currents in the irradiation area that would
not affect the fiber bundles could have a significant effect on the stability of the
coupling of the single fiber. Within such experimental variables, however, the
radiation response of the I-279 is similar to that of the low loss type B.
Figure 10. Change in Transmission Versus Dose, Corning I-279
-70
E -" ;;; -60 "0
~ -50 VI VI
i:i -40 z <!
~ -30
~
~ -20 z <! J: u -10
2
..
4 6
FIBER LENGTH: 22 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH:905nm
8 10 12 14 16 DOSE (KILORADS)
4. Wall, J. A. (1975) Transient Radiation Effects Tests of a Corning Radiation Resistant Fiber, AFCRL-TR-75-0012.
24
18
No annealing of the induced transmission loss was observed in the 1-279, but only
short periods of time (10-15 min) were allowed for observation of this effect. It
was not possible to have the fiber in position for several hours as in the case of
the type B to check for long-term annealing.
Figures 11 and 12 show the pre- and post-irradiation transmission spectra
for the low-loss types A and B respectively. As in the case of the commercial
glass fibers, both post-irradiation spectra show large increases in absorption at
the shorter wavelengths. The absorption induced in the type B, however, decreases
much more rapidly toward the longer wavelengths than does the absorption induced
in the type A. Although the type B is more radiation res istant than the type A
even at the shorter wavelengths, the difference between the two is significantly
enhanced in the near infrared because of this more rapid decrease in the induced
absorption.
3.2.3 PLASTIC FIBERS
Two types of plastic optical fibers were tested: Dupont Crofon@ 1110 and
International Fiber Optics 1032. The Dupont fiber has a polymethylmethacrylate
core and the International has a polystyrene core. Most, if not all, of the plastic
optical fibers available are made with one of these core materials. Tests of other
plastic fibers were therefore not considered necessary.
5000
4000
E "'" ..... CD
::'.3000 z o ~ :::> z w I-- 2000 !;!
1000
400
+-- AFTER DOSE OF 1000 RADS
600 800 1000 WAVELENGTH (nm)
1200
25
Figure 11. Attenuation Versus Wavelength, Corning Low Loss "A"
2200
2000
1800
1600
E -AFTER DOSE OF 1.6 x 105 RADS "" ..... 1400 !II
:3 z 0 1200 ~ :::l Z
1000 w 1-
~ 800
600
400
200
400
.' .
600 800 1000 1200
Figure 12. Attenuation Versus Wavelength, Corning Low Loss "B"
WAVELENGTH (nml
Figure 13. Change in Transmission Versus Dose, Dupont Crofon R 1110
-500
FIBER BUNDLE LENGTH:13.1 METERS -450 RADIATION: 10 MeV X-RAYS
SOURCE WAVELENGTH: 660nm
-400
E "" ;C -350 :3 z 2 -300 (J) (J)
~ -250 z <t 0:: 1- -200 ~ w <.!) -150 z <t J: U
-50
--INCREASING PULSE RATE __
BOMIN
10 MIN
20MIN
40MIN
50 MIN
20 HRS
Ok--~-~--~--L-~--~--L--L-~
o 10 20 30 40 50 60 70 80 100
DOSE (K1LORADSl
26
The test results for the Dupont Crofon® are shown in Figure 13. The time
intervals shown on the plot at different dos es are points at which the irradiation
was stopped to observe annealing of the induced transmission loss. The indicated
time intervals are the accumulated times following cessation of the irradiation.
For example, at 67 kilorad the first point below the maximum was taken 10 min
after the irradiation was stopped, the second point 20 min after the end of irradia
tion, etc. The data show that annealing of the induced transmission loss pro
ceeded fairly rapidly, sufficiently so that the results obtained were dependent on
the rate at which the linear accelerator was pulsed. Unfortunately, during the
irradiation this effect was not anticipated and an accurate record of the pulse
repetition rate was not kept. When the induced transmission loss appeared to be
changing too slowly, the pulse rate was increased until the signal amplitude began
to change at an acceptable rate. The maximum pulse repetition rate used, how
ever, was 60 pulses per sec and this applies in the region between 55 and 67
kilorad. The dose per pulse was 3.2 rads.
The test results for the International Fiber Optics fiber are shown in
Figure 14. In this case the relationship between the annealing rate and the
accelerator pulse rate was anticipated. As Figure 14 shows, the irradiation waS
continued at a fixed pulse rate until the induced transmission loss stopped
increasing with accumulated dose. The pulse rate was then increased and the
-500
-450 13PPS
-400
E ~ -350 III '0
~ -300 Ul Ul
~ -250 Ul z <t := -200
~
-50 10MIN
FIBER BUNDLE LENGTH: 5.1 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH: 660nm
39PPST 60PPS
/ /
/' /
/~ /'
5MIN
10 MIN 15 MIN
30 MIN
/ I
I I I I I I
/ 1 MIN
2 MIN
3 MIN 5 MIN
10MIN
15 MIN
('~ 65 HRS
O~~L-~~~~~~~~~~~~~-L~~
Figure 14. Change in Transmission Versus Dose, International Fiber Optics 1032
o 40 80 120 160 200 240 280 320 360 DOSE (KILORADS)
27
procedure repeated. Periodic stops of the irradiation to observe annealing are
indicated on the plot. For this irradiation the dose per pulse was 2.2 rads.
An interesting feature of the data shown in Figures 13 and 14 is that, following
each annealing period, the initial slopes, that is, rates of change in transmission
with dose of the curves are approximately the same regardless of the pulse repeti
tion rate. For both fibers the slope is about 0.02 dB/km.,.rad. This suggests that
the initially induced transmission loss is proportional to accumulated dose. How
ever, as is most obvious from the International Fiber Optics fiber data, the
equilibrium transmission loss attained is a function of pulse rate. The rate of
annealing of the induced transmission loss must therefore be a function of the in
duced loss. For example, referring to Figure 14, at a pulse rate of 13 pps equi
librium was reached at a loss of 165 dB/km. For equilibrium to occur the loss
induced by each pulse must have annealed out shortly before the next pulse ar
rived, in this case in 0.077 sec. At 39 pps, the loss induced by a single pulse
had to anneal out in 0.026 sec at the 365 dB/km level. Because of the strong de
pendence of annealing rate on induced loss, the results of these and other radiation
effects tests of plastic fibers must be examined with respect to the environment in
which the fiber is to be used. For a potential nuclear weapons environment,
transient radiation effects tests should yield more applicable data than the per
manent effects tes ts.
The long-term annealing data shown in Figures 13 and 14 can be used to esti
mate the extent of recovery of the induced transmission loss following the end of
an irradiation. The ratios of the losses observed at the indicated times after the
irradiation was stopped were calculated with respect to the induced loss observed
immediately at the end of the irradiation. A log-log plot of the ratios versus time
yielded a straight line, as shown in Figure 15, for the International Fiber Optics
data. The ratios (relative transmission losses) shown on the plot combine the
0:
:2 1.0 g z o ~ 0.5 ~ en z <t 0: f-w 0.2 > ~ ...J W
0: 0.1 ~1 ---2;;-------'--'--5::---'L...l......1....L:':10
c-----:2=-'=0c---L-..l--='"50,,--L...J
TIME AFTER IRRADIATION (t, minutes)
28
Figure 15. Relative Induced Transmission Loss Versus Time After Irradiation, International Fiber Optics 1032
annealing data taken at 127 kilorads and 347 kilorads. A least-squares fit to the
data gave the relationship:
R = 0.79 t -0.34 (2)
where R is the ratio' of the induced loss at time t (minutes) after the irradiation was
stopped to the loss observed before significant annealing occurred. Similarly, the
relationship
R=1.4t- 0• 3 (3)
was obtained from the Crofon® data
Equations 2 and 3 obviously cannot be applied to annealing times shorter than
those for which R = 1. For Eq. (2), t cannot be less than O. 5 min and for
Eq. (3) t cannot be less than 3 min. Also, (2) and (3) do not take into account
damage that does not anneal out and both give results that are low by ~pproximately
a factor of two for the 20 hr and 65 hr annealing times.
Figures 16 and 17 show the pre- and post-irradiation transmission spectra
for the Crofon® and International Fiber Optics fibers respectively. Within the
accuracy of the measurements, the Crofon® spectrum was essentially unchanged
by the irradiation except for a slight decrease in the intensity of the absorption
bands at wavelengths greater than 850 nm. ·A radiation induced increase in attenua
tion at wavelengths below 650 nm is the only significant change observed in the
International Fiber Optics transmission spectrum.
Figure 16. Attenuation Ver§1ls Wavelength, Dupont Crofon\.tll 1110
E -'" "-OJ .., '" 9 z 0 i= « ::J z lJJ ~
!;i
30
28
26
24
22
20
18
16
14
12
10
8
6
4 2
• BEFORE IRRADIATION->"
AFTER DOSE OF 6.7. 104 RADS~
400 600 800 WAVELENGTH (n m)
29
1000 1200
14
E 12 -'" iIi ."
'" 10 2 z o i= <t :J Z lLJ I--I--<t 4
2
AFTER DOSE OF 3.5 x 105 RADS
II'
Figure 17. Attenuation Versus Wavelength, International Fiber Optics 1032
400 600 800 1000 1200
WAVELENGTH (nm)
3.2.4 CERIUM DOPED FIBERS
Cerium doped optical fibers were specially prepared for this project by
Galileo Electro-Optics Corp. The base glass from which the fibers were drawn
was zinc crown. Three different glass melts doped with O. 1, 0.2, and 0.3 per
cent cerium were prepared and fibers drawn from them. The fibers were supplied
as 15 m long jacketed bundles each containing approximately 100 fibers. The in
trinsic attenuations of the fibers as determined by the manufacturer were 1250,
1200, and 1500 dB/km for the 0.1, 0.2, and 0.3 percent derium doping respect
ively.
The results of the tests of the cerium doped fibers are shown in Figure 18.
All three fibers are much more resistant to radiation induced transmission losses
than the commercial glass fibers and there is a significant increase in this resist
ance with increasing cerium doping. Comparison of the data for the 0.1 percent
and O. 2 percent cerium doped fibers shows that the extent to which the induced
losses annealed out also increased with increased cerium doping. Both fibers
were irradiated until the induced loss was about 400 dB/km. Sixty-eight hours
after the irradiation the 0.2 percent doped fiber had regained 225 dB/km of the
induced loss while the 0.1 percent doped fiber had regained only 100 dB/km 66 hr
after the irradiation. It is not clear from the data whether or not increasing
the cerium doping from 0.2 to 0.3 percent improved the degree of annealing be
cause the 0.3 percent doped fiber was not irradiated to as high an induced loss
as the other two fibers. However, the 0.3 percent fiber regained 200 dB/km of
the induced loss 67 hours after the irradiation, which is about the same as for the
0.2 percent doped fiber, suggesting that the additional cerium doping did not sig
nificantly affect the extent of annealing.
30
Figure 18. Change in Transmission Versus Dose, Cerium Doped Glass Fibers
4000
-450
-400
E -" -350 ..... ID "0
z -300 o (J)
!!! -250 ~ (J)
z ~-200 r-z - -150 w (!) z ~-100 u
-- AFTER DOSE OF 2300 RADS
E -" ..... ID "0
1000
BEFORE IRRADIATION
WAVELENGTH (nm)
31
.2%Ce
.3%Ce
BOMIN 30MIN
6BHRS
67HRS
FIBER BUNDLE LENGTHS: 13 METERS RADIATION: 10 MeV X-RAYS SOURCE WAVELENGTH: 905nm
DOSE (RADS)
Figure,19. Attenuation Versus Wavelength, O. 1 percent Cerium Doped Glass Fiber
3000
E -'" "-!D ~ Z 0
<- AFTER DOSE OF 5500 RADS
i= « :::J z W I-
~
1000 I
BEFORE IRRADIATION
WAVELENGTH (nm)
Figure 21. Attenuation Versus Wavelength, 0.3 percent Cerium Doped Glass Fiber
E -'" "-OJ ~ Z 0 i= 2000 « :::J z W I-
~
1000
600
32
Figure 20. Attenuation Versus Wavelength, 0.2 percent Cerium Doped Glass Fiber
<- AFTER DOSE OF 6900 RADS
800 1000 1200 WAVELENGTH (nm)
The slopes of the curves drawn through the data points in Figure 18 decrease
with increasing dose. As in the case of the Corning low.,.loss type B (paragraph
3.2), it cannot be certain whether the indicated nonlinear relationship between in
duced change in transmission and dose is characteristic of the fibers or is due to
rapid partial annealing of the induced losses between x-ray pulses.
Figures 19, 20, and 21 show the pre- and post-irradiation spectra for the
cerium doped fibers. The pre-irradiation spectra clearly show the absorption
bands resulting from the cerium doping at wavelengths below 700 nm. The post
irradiation spectra show the broadening of these bands caused by the trapping of
electrons generated during irradiation. At wavelengths greater than 900 nm the
separation between the pre- and post-irradiation spectra for each fiber remains
very nearly constant. This means that the radiation hardness of this type of fiber
is essentially independent of wavelength between 900 and (at least) 1200 nm.
3.2.5 RELATIVE SENSITIVITIES OF FIBERS TO X-RAYS
The initial slopes of the plots of radiation induced change in transmission
versus dose may be used to estimate the relative sensitivies of the fibers tested
to damage by x-rays. The slopes correspond to the rate of increase in loss of
transmission with increasing accumulated dose. The initial slopes were chosen
for this comparison because they start at "zero-dose", a common reference point
for all irradiations, and avoid the ambiguities involved in the nonlinear responses
observed with some of the fibers. Table 1 shows the slopes (initial rate of induced
loss) obtained from the fiber test results.
Table 1. Comparison of Sensitivities of Fibers to X-Radiation
Initial Rate of Induced Loss Fiber Type Fiber Des ignation (db/km-rad)
Bendix K2K 2.49 Commercial Glass Rank Hi-Tran 2.42
Schott LK 1. 67
Low-Loss Corning Low-Loss "A" 0.95
Fused Silica Corning Low-Loss "B" 3.4 X 10-3
Corning 1279 5.6 X 10-3
Dupont Crofon® 0.055 Plastic Fibers International Fiber
Optics 1032 0.035
Cerium Doped O. 1 percent Ce 0.2 0.2 percent Ce 0.084
Fibers O. 3 percent Ce 0.079
33
Comparison of the initial rates of induced loss is obviously not a perfect
measure of the relative radiation sens itivities of optical fibers. For example,
Figure 18 shows the 0.3 percent <:erium doped fiber to be much less sensitive to
radiation compared to the 0.2 percent cerium doped fiber than the rates of induced
loss shown in Table 1 imply. However, Table 1 clearly shows that the germanium
doped fused silica fibers (Corning Low-Loss "B" and !-279) are the least sensitive
to x-radiation of the fibers tested. These are followed by the plastic fibers, the
cerium doped fibers, the titanium doped fused silica fiber (Corning Low-Loss "A"),
and the commercial glass fibers in order of increasing radiation sensitivity. If
fast annealing of induced loss is considered an important factor in determining the
radiation hardness of a fiber, the plastic fibers would be chosen as the least sensi
tive to radiation. Such information is not shown in Table 1 and must be obtained by
reference to the original data.
3.3 Neutron Effects
3.3.1 COMMERCIAL GRADE GLASS FIBERS
Two commercial grade glass fibers were tested for their responses to
neutrons. These were Bendix K2K and Corning 5011 fibers. The results of the
tests are shown in Figures 22 and 23 respectively. For these and other
E -" "-m ~ z 0 CJ) CJ)
:::;: CJ)
z <i cr I-
:;;: IJ.J (!) z <i :I: u
-900
-800
-700
-600
-500
-400
-300
-200
ACCUMULATED GAMMA DOSE (RADS)
100 200 300 400 500 600 700
FIBER BUNDLE LENGTH: 13.4 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 905nm
°0L-~~~~--.8L-L-~l.2~--lL.6-L-2~D~--2LA-L-2~.-8~-3L.2~
ACCUMULATED NEUTRON FLUENCE (l012n/cm2)
34
Figure 22. Change in Transmission Versus Neutron Fluence, Bendix K2K
Figure 23. Change in Transmission Versus Neutron Fluence, Corning 5011
E -'" ..... (l] "C
z 0 en en ::E en z « 0:: f-
~ lJJ (!)
z « :r: u
-1100
-1000
-900
-800
-700
-600
-500
-400
-300
-200
-100
0 0
ACCUMULATED DOSE (RADS) 100 200 300 500 600 700
.4 .8
FIBER BUNDLE LENGTH': 13.4 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 905 nm
1.2 1.6 2.0 2.4 2.8 3.2
ACCUMULATED NEUTRON FLUENCE (J012n/cm2
)
representations of the neutron effects data, the data point at the upper end of
each line segment represents the change in signal transmission measured im
mediately following a reactor burst. The data points below this represent the
change in transmission measured at the indicated times following the burst. The
lower abscissa gives the sum of the neutron fluences at the fiber for all bursts up
to the indicated data points and the upper abs cissa is the corresponding sum of the
gamma doses in rad (Si) for the bursts.
For the Bendix K2K data (Figure 22) the average slope, relative to accum
ulated gamma dose, of the lines drawn between data points is 2.5 dB/km-rad.
This is almost identical to the initial rate of induced loss (2.49 dB/km-rad) found
in the x-ray effects tests (see Table 1, paragraph 3.2.5). It may therefore be
concluded that the induced losses observed were produced primarily by the gamma
radiation associated with the reactor bursts and that the neutrons caused relatively
little change in transmission. The neutrons apparently had a significant affect on
the annealing of induced loss, however. No short term annealing was observed
during the permanent x-ray effects tests of this fiber, although 40 days after the
first irradiation the induced change in transmission dropped from 114 to 62 dB/km
(see Figure 2, paragraph 3.2.1). The data in Figure 22 show much more rapid
35
annealing. For example, after an accumulated neutron .fluence of 3.3 X 1011
neutrons/cm2 and gamma dose of 74 rads, the induced change in transmission
dropped from 197 to 115 dB/km in 30 min.
The data for the Corning 5011 shown in Figure 23 is similar in form to the
Bendix K2K data. Since no permanent x-ray effects tests were performed on the
5011, a comparison of the effectiveness of the gamma radiation relative to the
neutrons in producing damage in the fiber is not possible. Neither can any con
clusions be drawn regarding possible neutron-induced increases in annealing rate.
However, it was found in the transient x-ray effects tests that the rate of induced
loss for the 5011 was about 40 percent higher than for the K2K. The average slope,
relative to gamma dose, of the lines between data points in Figure 23 is 3.7 dB/
km-rad. Increasing the average slope for the K2K data by 40 percent gives
3.5 dB/km-rad. The closeness of these two values makes it reasonable to con
clude that the relative gamma-neutron response of the Corning 5011 is similar to
that of the Bendix K2K.
3. 3. 2 LOW - LOSS FUSED SILICA FIBERS
The Corning low-loss type Band 1-279 germanium doped fused silica fibers
were tested. The 1-279 single fiber was exposed to three successive reactor bursts,
each producing a neutron .fluence of 1. 0 X 10 12 n/ cm 2 and a gamma dos e of 217 rad
at the fiber position. No change in the amplitude of the signal transmitted by the
fiber was observed immediately following any of the bursts. The amplitude of the
signal from the photodetector was low, however, and it is possible that small
changes in amplitude occurred that could not be measured.
The results of the low-loss type B irradiation are shown in Figure 24. Mal
functions of the oscilloscope camera prevented measurement of the signal ampli
tude immediately following the second and fourth reactor bursts. The delays of
3 min and 1 min that occurred before the signal was photographed are indicated
in Figure 24. The data at 1. 2 X 1011 neutrons/cm2, taken after the first burst,
show that significant annealing probably occurred during these delay times.
Also, 45 min elapsed between the second and third bursts and no photograph of the
signal was obtained to determine the amount of annealing that could have occurred
during this period. Dashed lines are shown connecting these data points to in
dicate that they should not be used to determine rates of induced loss.
In spite of these problems, the data shown in Figure 24 can be used to draw
some conclusion about the effects of neutrons on the low-loss type B fiber. The
slope, relative to gamma dose, of the line drawn from the origin to the first data
point at 1. 2 X 1011 n/ cm 2 and 31 rads is O. 98 dB /km - rad. This is cons iderably
greater than the initial rate of induced loss of 3.4 X 10-3 dB/km-rad determined
in the permanent x-ray effects tests of this fiber. It may be concluded from this
36
Figure 24. Change in Transmission Versus Neutron F1uence, Corning Low-Loss "B"
-100
-90
-80
E x m -70 -0
~ -60 iii If)
~ -50 If)
z « 1= -40
~ -30-
ACCUMULATED GAMMA DOSE (RADS)
100 200 300 400 500 600 700
" "
FI BER BUNDLE LENGTH: 14 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 905 nm
" I,
SEE TEXT / ' A (1 MIN) '\ I ' / I \.,'; // J
I I ;' I t I ,,/ I
I I " I / ,/ I
I I ,,/ I
/(3MIN) ~/30MIN 130M'N , , 'j'-I MIN
z « :r: u
-20 - ~~O MIN
-10 .
.4 .8 1.2 1.6 2.0 2.4 2.8 3.2
ACCUMULATED NEUTRON FLUENCE (l0'2n/cm2)
that, unlike the commercial glass fibers, the neutrons rather than gamma radia
tion are the primary cause of induced change in transmission for this fiber.
The data point at 9 X 1011 n/cm2 and an induced change in transmission of
65 dB/km was taken immediately following a reactor burst. Thirty minutes later
the transmission had improved by over 34 dB/km. Judging from the data at
1. 2 X 1011 n/cm2, most of this annealing probably occurred during the first few
minutes following the burst. Referring to the permanent x-ray effects data for the
low-loss type B shown in Figure 9, paragraph 3.2.2, after a dose of 20 kilorads
and an induced change in transmission of 58 dB/km, the transmission improved by
only about 8 dB /km 15 min after the irradiation was s topped. The annealing rate
of transmission losses induced by the neutron irradiation is therefore much greater
than that of the x-ray induced losses.
3.3.3 PLASTIC FIBERS
Only the Dupont Crofon® was tested for its response to neutrons. The results
of the irradiations are shown in Figure 25. The slope relative to gamma dose of
the line from the origin to the data point at 5.45 X 1011 n/cm2 and 120 rads is
0.14 dB/km-rad; for the line to the point at 1. 66 X 1012 n/cm2 the slope is
0.18 dB/km-rad. These slopes are roughly three times the initial induced loss
37
ACCUMULATED GAMMA DOSE (RADS)
_90.-,-_1~O~O-,~20rO~.-3~OrO~~4TO~0-,~50rO~~6~0~0~
-80
]" -70 .... CD
" -60 z o (/) (/)
::; (/)
z <{ a: f-
-50
-40
~ -30 w
FIBER BUNDLE LENGTH: 13.4 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 660 nm
I I I I I I I
! :::r /j """ '"'' O~:" , " ,-,
I 2MIN OR 30MIN
_L _.~L ~ __
o .4 .8 1.2 1.6 2.0 2.4 2.8 3.2
ACCUMULATED NEUTRON FLUENCE (lO'2n/cm2
)
Figure 25. Change in Transmission VersuANeutron Fluence, Dupont Crofon<!9 1110
(0.055 dB/km-rad) found for this fiber in the permanent x-ray effects tests. The
observed changes in transmission were therefore caused by both the neutrons and
the gamma radiation, with the neutrons producing most (60 to 70 percent) of the
changes.
It was found in the permanent x-ray tests of this fiber that the annealing rate
of the induced loss is a function of the loss. Since the changes in transmis~ion
produced during the neutron irradiations were less than those for which annealing
was observed in the permanent x-ray tests, a direct comparison of annealing
rates between the two tests is not possible. Annealing of the losses induced in the
Crofon® by the reactor radiation is unusual in that it apparently reached its limit
within 2 min following a reactor burst. This is indicated in Figure 25 which shows
that the change in transmission measured 2 min after a reactor burst was the
same as that measured 30 min later. It cannot necessarily be conci uded from
this, however, that the annealing rate following neutron irradiation is greater than
that following irradiation with x-rays.
3.3.4 CERIUM DOPED FIBERS
Three cerium doped glass fibers, identical to those used for the permanent
x-ray effects tests, were tested. Some anomalous results, indicative of erratic
optical coupling between the fiber and either the LED light source or photodetector,
were found when the 0.2 percent cerium doped fiber data was analyzed. Therefore,
38
fore, only the results obtained for the 0.1 and 0.3 percent doped fibers are pre
sented here. These are shown in Figures 26 and 27 respectively.
For the O. 1 percent cerium doped fiber, the average slope, relative to
gamma dose, of the lines drawn between the data points in Figure 26 is
0.59 dB/km-rad. This is three times the initial rate of induced loss (0.2
dB/km-rad) found in the permanent x-ray effects tests of this fiber. Similarly,
for the 0.3 percent doped fiber the average slope of the lines in Figure 27 is
0.23 which is also about three times the initial rate of induced loss (0.079
dB / km -rad) found in the x -ray tes ts. It can therefore be concluded that, like the
plastic fiber, the changes in transmission observed in these fibers were due
to both the neutrons and gamma radiations, with 60 to 70 percent of the changes
being caused by the neutrons.
Although the data show that the 0.3 percent cerium doped fiber is significantly
less sensitive to losses induced by neutrons than the 0.1 percent doped fiber, no
specific conclusions can be drawn from the data regarding the relative rates of
annealing of the induced losses. Also, the changes in transmission produced in
the fibers by the reactor radiations were less than those for which annealing was
observed in the permanent x-ray tests, so it is not possible to make a valid com
parison between annealing results for the neutron and x-ray tests.
Figure 26. Change in Transmission Versus Neutron Fluence, O. 1 percent Cerium Doped Glass Fiber
ACCUMULATED GAMMA DOSE (RADS)
100 200 300 400 500 600 -180r--'~~'--,~~,,~,-~-,~~-,--~
-160
~ -140 ..... (D
" z -120 o (J) (J)
::; (J)
z « a:: I--
-100
-80
~ -60 w (!) z ~ -40 u
39
I I I I I 130MIN
FIBER BUNDLE LENGTH: 13.4 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 905 nm
ACCUMULATED NEUTRON FLUENCE (I012n/cm2
)
ACCUMULATED GAMMA DOSE (RADS)
-18 ° r----r--,-I 0[-,0=--.,-,2=-;0,-,0'-,--,3,-;0-,,0'-,--,4,-;0-,,0'-,--,,5,-;0-,,0-,--,,6-:;0-=-0-,---,-7-;-0.:,0
-160
E -140 oX "CD
" ~-120 z Q
~ -100 ::< en
FIBER BUNDLE LENGTH:13.4 METERS RADIATION: WHITE SANDS FBR SOURCE WAVELENGTH: 905nm
~ -80 a: I-
~ -60 w
, I I I
Figure 27. Change in Transmission Versus Neutron Fluence, 0.3 percent Cerium Doped Glass Fiber
<!l
~ -40 12MIN I
J: u
I I I 117HRS
.4 .8 1.2 1.6 2.0 2.4 2.8 3.2.
ACCUMULATED NEUTRON FLUENCE (I012n/cm2)
3.3.5 RELATIVE SENSITIVITIES OF FIBERS TO NEUTRONS
If we assume that the initial rates of induced loss found for x-rays can be
applied linearly to the gamma radiation that accompanied the neutrons during
the reactor irradiations of the fibers, the relative sensitivities of the fibers
to neutron induced losses can be estimated. Using this assumption the change
in transmission caused by the gamma radiation alone is given by:
.6T rD 'Y
where
.6T the change in transmission caused by gamma radiation 'Y
r initial rate of induced loss for x-rays as given in Table I
paragraph 3.2.5
D = the dose due to gamma radiation during a reactor burst.
The change in transmission caused by the neutrons alone is then:
.6Tn = .6T - rD n'Y
40
(4)
(5)
where
.6T = the change in transmission caused by neutrons n
.6T = the change in transmission produced by the mixed neutron-gamma n')'
radiation.
Dividing .6Tn by <1>, the neutron fluence in units of 1012 n/cm2 for a given
reactor burst, gives the change in transmission produced in a fiber per unit neutron
fluence. These calculations were performed for the first reactor burst to which
each fiber was exposed and the results are shown in Table 2. In Table 2 .6T / <1>
is in units of dB/km per 1012 neutrons/cm2. Results for the Corning 5011 fiber
are not given in the table because no initial rate of induced loss for x-rays had
been measured for it.
Table 2. Comparison of Sensitivities of Fibers to Neutrons
Fiber Type Fiber Des ignation Neutron Sensitivity
.6T / <1>
Commercial Glass Bendix K2K 161
Low-Loss Corning Low- Loss "B" 254
Fused Silica Corning 1-279 -0
Plastic Dupont Crofon ® 19.9
O. 1 percent Ce 79.2 Cerium Doped 0.3 percent Ce 44.7
With the exception of the Corning low-loss type B, the different types of fiber
are in the same order of sensitivity to neutron induced losses as for x-ray induced
losses. The high value of .6T/<1> obtained for the low-loss type B fiber is difficult
to accept, particularly in view of the fact that a similar fiber like the Corning 1-279
showed no detectable change in transmission when exposed to a greater neutron
fluence (paragraph 3.3.2). Since there is nothing in the data to justify rejecting
this high value of .6T / <1>, one can only speculate regarding the reasons for it.
Table 2 is presented solely for the purpose of comparing the measured neutron
responses of the fibers tested, and no additional implications should be applied
to the listed values of .6T / <1>. Specifically this method used to calculate .6T /<1> is
not intended to imply that the complex effects, produced by the simultaneous
irradiation of the fibers by neutrons and gamma rays, can be easily separated.
41
Figure 28. Signal Transmitted by 4.5 m Length of Dupont Crofon ® Plastic Fiber Irradiated With 1200 rad, 20 nsec X-ray Pulse
3.4 Transient X-Ray Effects
The different types of fibers tested
all had a common form of response
when subjected to intense x-ray pulses.
The form of this response is illustrated
in Figure 28 which is a photograph of
the oscilloscope trace of the signal
transmitted by a 4. 5 m length (7.5 m
including leads) of Dupont Crofon®
plas tic fiber when it was irradiated by
a 1200 rad, 20 nsec x-ray pulse. For
this irradiation the oscilloscope trace
position was set two major divisions
(centimeters) below the graticle
centerline with no input signal. The
signal amplitude from the photo
detector prior to irradiation was
105 mv, putting the oscillos cope trace
1 mm above the centerline. The
photograph shows a pulse, greater in
amplitude than the pre-irradiation
signal, at the beginning of the trace.
This pulse is due to fluorescent radiation produced in the fiber by the x-rays.
Following the fluorescence pulse the signal amplitude has dropped to about 50 per
cent of its original value, showing a radiation induced change in the transmission
of the fiber. The signal level, and therefore the transmission, then approaches
its pre -irradiation value over a period of 500 J.!s ec.
Because all of the fibers tested showed both a radiation induced transient
change in transmission and a fluorescent pulse, the results of the transient x-ray
tests are classified according to these two characteristics instead of according to
fiber types as was done in previous sections.
3: 4. 1 TRANSIENT CHANGE IN TRANSMISSION
The initial magnitude, rate of recovery, and degree of recovery of the
transient change in transmission differed considerably among the types of fibers
tested. To characterize these differences, the amplitude of the signal transmitted
as a function of time following an x-ray pulse was determined for each fiber from
oscilloscope photographs similar to that shown in Figure 28. Using the signal
amplitude measured before irradiation and applying Eq. (1), paragraph 3. 1, a
table of values of the induced change in transmission in dB/km versus time was
42
obtained for each fiber. These results were then normalized to the dose received
by the fibers during irradiation. It was found possible to fit these data to within
10 percent with a relation of the form"~
where
~T = induced change in transmission per unit dose (dB/km-rad)
= time after x-ray pulse (milliseconds).
A, B, C, 71 and 72 are constants characteristic of the fiber.
(6)
The characteristic constants found for the fibers tested are shown in Table 3.
Table 3. Characteristic Constants of Eq. (6) for Transient Radiation Induced Change in Transmission of Fibers
Fiber Fiber Type Designation
Commercial Bendix K2K
Glass Corning 5011 (note)
Corning Low-loss Low-loss "A" Fused Silica Corning 1-279
Plastic Dupont Crofon®
1. F. O. 1032
0.1 Percent Ce Cerium 0.2 Percent Ce Doped
0.3 Percent Ce
Units: Dose in Rads (SO
A, B, C in dB/km-rad
71' 72 in milliseconds
Dose A
93 7.17
92 11. 1
230 1. 03
1000 0.161
1200 0.259
940 0.167
120 0.633
160 0.441
150 0.339
B C 71 72
6.77 3.95 18.9 0.203
- 5.51 2.49 -
1. 04 0.605 27.0 0.397
0.213 0 0.658 0.041
0.347 0 0.338 0.382
0 0 0.0223 0
0.460 0.156 16.1 0.306
0.276 0.0648 11. 3 0.266
0.389 0.032 15.4 0.306
Note: Insufficient data for complete fit. Applies only for 7 .:::. O. 5 msec.
'~Attempts were made to fit the transient data with functions having the form At- a which was found to apply to the long term recovery of induced losses for the plastic fibers (paragraph 3.2.3), but it was found that this type of function could be applied only to fairly limited ranges of time.
43
The values of the constants shown in Table 3 were obtained from least squares
fits to data covering periods up to 10-30 msec after the x-ray pulses. For most
of the fibers tested the induced changes in transmission continued to anneal out
over much longer time periods. This long-term annealing, similar to that ob
served in the permanent x-ray effects tests, was not considered appropriate for
inclus ion in the trans ient radiation effects tes t res ults. Equation (6), along with
the given constants, is therefore not necessarily representative of the transient
responses of the fibers for times greater than the order of 10 msec after irradiation
with an x-ray pulse.
The doses listed in Table 3 are those received by the fibers from a single
x-ray pulse during the tests. These are the doses used to normalize the measured
changes in transmission before fitting the data to Eq. (6). The normalization was
performed in order to make a comparison of the transient responses of the dif
ferent fibers possible even though they had not received the same doses. (Higher
doses were required for some of the fibers in order to obtain changes in signal
amplitude sufficient for accurate measurements to be made). This procedure
implies that the induced changes in transmission are proportional to dose. Al
though tests were performed with the Corning low-loss type A fiber at doses of
30, 230, and 700 rads that showed this proportionality, it does not necessarily hold
for all fibers or over all dose ranges. However, the assumption of proportionality
between dose and induced change in transmission can be eliminated from the in
formation given in the table by multiplying the constants A, B, and C by the ap
propriate dose. Applying Eq. (6) will then give the changes in transmission, in
dB/km, actually observed during the tests.
The magnitude of the initially induced change in transmission and the rate at
which the induced change recovers may be used as criteria for ranking the fibers
according to their res is tance to trans ient radiation effects. The s urn of the
constants A, B, and C gives a measure of the initially induced change in trans
mission. Based on this criterion alone, Table 3 shows the International Fiber
Optics 1032 plas tic fiber to be the mos t res is tant to trans ient effects. This is
followed by the CorningI-279 (germanium doped fused silica), the Dupont Grofon@
(plastic), the cerium doped fibers, the Corning low-loss type A (titanium doped
fused silica), and the commercial glass fibers. The smaller the constants 71
and 72' the faster the rate of recovery of the induced change in transmission.
Bas ed on the tabulated values of 71 and 72, ignoring the Corning 5011 results
which were bas ed on a limited amount of recovery time data, the fibers rank in
the same order of radiation resistance except that the Dupont Crofon R shows a
faster recovery time than the Corning I-279.
44
3.4.2 FLUORESCENT PULSE
Figure 29 shows an oscilloscope
photograph of the signal produced by the
fluorescent pulse generated in a Corning
1-279 single fiber. For this photograph,
a 9.8 m length (total length 11. 4 m in
cluding leads) of the fiber was exposed to
a 1000 rad, 20 nsec x-ray pulse. There
was no light input to the fiber and the
zero-signal position of the oscilloscope
trace was set one major division (centi
meter) below the graticle centerline.
The fluorescence signal has its maxi
mum amplitude at the beginning of the
trace, corresponding to the incidence of
the x-ray pulse. The amplitude decreases
exponentially, as determined by a least
squares fit to data taken from the photo
graph, toward the zero signal level
during a period of over 2/.lsec.
The Corning 1-279 single fiber is
the only fiber tested for which it was
poss ible to obtain complete temporal
Figure 29. Signal Produced by Fluorescent Pulse Generated in 9.8 m Length of Corning 1-279 Single Fiber Irradiated With 1000 rad, 20 nsec X-ray Pulse
data on the x-ray generated fluorescent pulse. The fluorescence produced in the
other fibers (which were tested as jacketed bundles containing many single fibers)
was of sufficient magnitude to overload the photodetector even at the lowest dose
(20 rads per pulse) available from the .flash x-ray machine. However, it was pos
sible to obtain some information on the fluorescence produced in the other fibers
by measuring the dynamic overload characteristics of the photodetector and com
paring thes e with the signals produced by the fluores cence during the irradiations.
The overload characteristics of the photodetector were measured by using a laser
diode that produced a 100 nsec wide pulse in combination with neutral density
filters to generate a wide range of signals from the photodetector. It was found
that as the intensity of the light input to the photodetector was increased, the output
signal amplitude increased linearly up to a maximum beyond which only the width
of the output signal increased. A direct correlation was found between the width of
the photodetector output signal and the intens ity of the light input and this was
used to determine the amplitudes of the fluorescent pulses obtained from the fibers.
The changes in the photo detector output signal decay time under overload conditions
were also measured and the results used to correct the observed decay times of
the .fluores cent puIs es.
45
The effective radiant intensity of the fluorescence emitted at the ends of the
fibers was estimated from the properties of the photodetector (Texas Instruments
TlXL79). The detector had a responsivity of 2 X 10 5 volts/watt at 900 nm and a
sensitive area of 4.6 X 10-3 cm2
• The distance from the detector window, against
which the fibers were butted, to the sensitive area was 0.19 cm giving a solid angle
of O. 127 sr. The effective intensity at 900 nm was therefore obtained from:
1= V = 3.94 X 10- 5 V 2 X 10 5 X O. 127
(7)
where
I = intensity in microwatts /steradian
and
V = photodetector signal amplitude in microvolts
In order to compare the intensities of the fluorescence generated in the dif
ferent fibers, the following normalization procedure was performed. Let i be the
fluorescent intensity generated per unit length and trapped in the fiber. Assuming
exponential absorption of the fluorescence radiation, the observed intensity at the
end of the fiber is:
f L . L . -aX 1 -a 1= le dx=-(1-e ) o a (8 )
where
I intens ity at end of fiber
intens ity generated and trapped per unit length of fiber
a absorption coefficient
L = length of fiber
x = distance from end of fiber at which fluorescence is generated.
From (8) we obtain
i 10' (9)
which normalizes the observed fluorescent intensity to fiber length. The value of
"all is directly related to the attenuation of the fiber expressed in dB/km. If "0 1
is expressed in meters
46
-4 a = 2.3 X Hi A (10)
where
A = fiber attenuation in dB/km
a = absorption coefficient in m- 1
In evaluating" a", the approximate intrinsic attenuations of the fibers were used.
This presumes that no radiation induced losses in the transmissions of the fibers
coincided with the generation of the fluorescent pulses.
It was found that the intensity of the fluorescence generated in the plastic
fibers was a linear function of the dose rate to which they were exposed. Assuming
this was also true for the glass fibers, the values of" i" obtained from Eq. (9)
were divided by the dose rates to which the fibers were exposed. An additional
division by the numbers of fibers in the irradiated bundles completed the normal
ization process, giving:
i n
where
Ill' -aL •
(l-e )D.N
in fluorescent intensity per unit length and dose rate per fiber
D dose rate
N number of fibers in bundle
(11)
Table 4 lists in for the fibers tested along with the values of the parameters
us ed in the calculations. Als 0 lis ted in Table 4 are the decay times constants
"'T" for the fluorescent pulses. Here it is assumed that the fluorescence decays
as:
I
where
10 = initial intensity of fluorescence
= time after x-ray pulse.
(12)
The number of fibers per bundle (N) listed in Table 4 were estimated from
the manufacturers I specifications of fiber and bundle diameters. Where N is
shown as 100 or more the numbers were rounded off to the nearest 100. All of
the fibers in a bundle were not necessarily continuous throughout the length of
47
,
Table 4. Fluorescent Pulse Characteristics
Fiber Fiber De- i I D Type signation n T
- - -
Bendix 5.3X10- 2 3.1 2.8X10 5 4.65X103 Com- K2K mercial Corning 4.6X10-3 5.1X104 4.6X103 Glass 5011 2.6
Corning 3.6X10-4 2. 1X102 1. 35X103
Fused low-loss 0.4
Silica "A" Corning 6,.9X10- 6 0.38 3.3 5X104 1-279
Dupont® Crofon
9.7X10-4 -0 1. 3X102 3.3X103
Plastic I.F.O. 1. 2X10- 3
1. 4X102 1. 65X103
1032 0.3
0.1 Per- l. 5X10-3 6.8 3.3X103 6. 15X103
Cerium cent Ce
Doped 0.2 Per- 2.2X10-3 7.8 6.2X103 8.05X103 Glass cent Ce
0.3 Per-l. 5X10-3 3 7. 65X103 cent Ce 6.1 3.3X10
Units - -
i pW /steradian-megarad/sec-m-fiber n
T psec
I pW /steradian
D megarads/sec
L m
A dB/km
L
4.7
5.2
4.6
9.8
4.5
4.5
13.4
13.1
13.4
N A
400 1000
800 1000
100 100
1 10
16, 1200
32 1500
100 1200
100 1200
100 1500
the bundle. A particularly extreme example of this is the Corning low-loss
type A, which is a fairly brittle fiber and eas ily broken. It had only about 20
percent of the fibers conducting light through the length of the bundle. Since
fluorescence is generated throughout the fibers regardless of whether or not they
are broken, the effects of incomplete continuity of a bundle could not be reasonably
accounted for. All of the fibers in a bundle were therefore assumed to be con
ducting when i was computed. The lengths of fiber (L) shown in the table are n the lengths actually exposed to the x-ray pulses. The total fiber lengths used in
the tests were about 3 m longer, the additional fiber being used as leads between
48
the irradiated coil and the photodetector and source. These leads were assumed
lossless when calculating in'
Except in the case of the Corning I-279 fiber, the values of I (and therefore
in) and 7" shown in Table 4 are only approximate. This is because of the cor
rections required in the data to account for the overloading of the photodetector.
Assuming the values of in are valid within a factor of two~ all of the fibers
apparently emit roughly the same amount of fluorescent radiation per unit length
and incident dose rate with the exceptions of the Bendix K2K and the fused silica
fibers. No explanation can be given for the much larger fluorescence observed
from the Bendix fiber than from the others. The lower normalized fluorescent
intensity emitted from the fused silica fibers, however, could be explained by the
fact that the optical acceptance angle for this type of fiber is four to five times
smaller than the acceptance angles of the other fibers. Fluorescence radiation
generated in the fused silica fibers is therefore more likely to escape through the
fiber surfaces rather than being conducted along the fiber. The much greater
fluorescent intensity observed from the Corning low-loss type A fiber bundle than
from the Corning I-279 single fiber could be the result of fluorescence radiation
escaping through the surfaces of the fibers and being pickup-up at the proper ac
ceptance angle by other fibers in the bundle. This greater" collection efficiency"
of the fiber bundle compared to the single fiber would result in an apparently
higher level of fluorescence per fiber.
The decay time constants of the fluorescent pulses for the fused silica and
plastic fibers are an order of magnitude shorter than for the commercial and
cerium doped glass fibers. Therefore, even if all the different types of fiber
emitted the same intensity of fluorescence, the fused silica and plastic fibers would
be preferable for use in a transient radiation environment because of the much
shorter duration of the fluorescent pulses that might be produced in them.
3.5 Summary of Results
To review the significance of the test results, they will be applied to estimate
the effects of radiation on 30 m lengths of the fibers tested. Table 5 shows the
doses required to reduce the transmissions of the fibers to 50 percent of their
pre-irradiation values; that is, an induced change in transmission of -100 dB/km
for a 30 m length. The results shown in Table 5 do not take into account annealing
of the induced losses and are taken directly from plots of the permanent x-ray
effects test data.
These results clearly show the superior resistance of the germanium doped
fused silica fiber (Corning low-loss type B) to x-radiation induced losses in trans
mission.
49
Table 5. X-Ray Dose Required to Reduce the Transmission of a 30 m Length of Optical Fiber to 50 Percent of its Original Value
Fiber Dose (Rads)
Bendix K2K 38
Rank Hi-Tran 38
Schott LK 65
Corning Low-Loss "A" 105
O. 1 Percent Ce Doped 500
0.2 Percent Ce Doped 1100
O. 3 Percent Ce Doped 1500
Dupont Crofon@ 5000
I.F.O.I032 8000
Corning Low-Loss "B" 40,000
Table 6 shows the estimated neutron fluences required to reduce the trans
missions of 30 m lengths of fiber to 50 pen:ent of the pre-irradiation values. The
results shown in this table were dex'ived from the neutron sensitivities (b.T / iJ? )
shown in Table 2, paragraph 3. 2. 5, and are irrespective of annealing effects.
Table 6. Neutron Fluence Required to Reduce the Transmission of a 30 m Length of Optical Fiber to 50 percent of its Original Value
Neutron Fluence
Fiber 12 2 (10 neutrons/cm)
Corning 0.39
Bendix K2K 0.62
0.1 Percent Ce Doped 1.3
0.3 Percent Ce Doped 2.2
Corning I-279 > 3
Dupont Crofon 5
50
As discussed in paragraph 3.2.2, there is a conflict in the neutron test
results obtained for the two germanium doped fused silica fibers. The Corning
low-loss type B fiber was found to be the most sensitive to neutron damage, as
Table 6 shows, while the Corning 1-279 fiber showed no detectable change in
transmission after exposure to 3 X 1012 neutrons/cm2• Because this difference
in neutron sensitivities could not be explained except on the basis of a possible
lower sensitivity for the Corning 1-279 tests, the Dupont Crofon® plastic fiber
is shown in the table as the least sensitive to neutron induced transmission losses.
Table 7. shows the times required for the transmissions of 30 m lengths of
fiber to recover to 50 percent of their original values after exposure to a 1000 rad
pulse of x-rays.
The recovery times were computed from Eq. (6) and the constants listed in
Table 3, of paragraph 3.3.1. Those fibers for which the recovery times would
exceed one min are not included in Table 7.
Table 7. Time Required for the Transmission of a 30 m Length of Fiber to Recover to 50 percent of its Original Value Following a 1000 Rad X-Ray Pulse
Recovery Time Fiber (milliseconds)
0.2 Percent Ce Doped 29
0.3 Percent Ce Doped 25
Dupont Crofon@ 0.32
Corning 1-279 0.31
r. F. O. 1032 0.011
The International Fiber Optics 1032 plastic fiber has a recovery time that is
shorter by more than an order of magnitude than its closest competitors, the
Dupont Crofon@ plastic fiber and the Corning 1-279 germanium doped fused silica
fiber. The 0.1 percent cerium doped glass fiber was excluded from the Table 7
because of its long recovery time. This, along with the results presented for the
other cerium doped fibers, shows that cerium doping can improve the transient
radiation effects properties oJ glass fibers as well as their resistance to per
manent radiation induced transmission losses.
Equation (11) in paragraph 3.3.2, was solved for the fluorescent radiant
intens ity per unit dos e rate ·per fiber (lIN, b) that would be emitted at the ends
of 30 m lengths of the fibers when exposed to x-ray pulses. The data shown in
Table 4, paragraph 3.3.2, were used to evaluate I/N' D and the results are
51
Table 8. Fluorescent Radiant Intensity at the End of a 30 m Length of Fiber Exposed to anX-Ray Pulse
.,. Fiber l/N.D (J,fsec)
Bendix K2K 0.23 3.1
Corning 5011 0.02 2.6
0.2 Percent. Ce Doped 8X10- 3 7.8
Corning Low-Loss "A" 7.8X10- 3 0.4
O. 1 Percent Ce Doped 5.4X10- 3 6.8
0.3 Percent Ce Doped 4.3X10- 3 6.1
1. F. O. 1032 3.5X10-3 0.3
Dupont Crofon ® 3.5X10-3 -
-0 --
Corning 1-279 2X10- 4 0.38
Units of l/N'I> : J,fW/steradian-megarad/sec-fiber
listed in Table 8. The fluorescent pulse decay times (.,.) listed in Table 4 are
also shown here for comparison purposes. Comparison of the values of l/N. I>
in Table 8 with the values of in -the fluorescent intensity per fiber per unit length
and dose-rate-in Table 4 shows that the length of fiber considered has a signi
ficant effect on the relative intensity of the fluorescent radiation emitted from
the end of the fiber. For example, the ratio of the value of i for the
Dupont Crofon@ to that for the Corning 1-279 is 141, while t~ ratio of the values
of l/N. D for the 30 m lengths is 17. 5. This is because the greater attenuation of
the Crofon@ reduces the amount of fluorescent radiation reaching the end of the
fiber, compensating for the higher intensity of fluorescence generated and trapped
per unit length. The lower intrinsic fluorescent emission of the fused silica fibers
is therefore negated to some degree in long lengths of fiber because of their
lower intrins ic attenuations.
The fused silica and plastics fibers have the shortest fluorescent pulse decay
time constants. Table 8 shows that the decay time constants are dependent only
on the type of fiber and do not correlate with the intensity of fluorescence.
In summarizing test results, germanium doped fused silica fibers appear to
be the most radiation resistant of the fibers tested. The next most resistant to
radiation damage are the plastic fibers, which have some advantages in faster
recovery of transient induced transmission losses and lower neutron induced
losses. The cerium doped glass fibers are more than an order of magnitude more
resistant to radiation induced losses than the commercial glass fibers. The latter
are very radiation sensitive and essentially unusable in radiation environments.
52
4. CONCLUSIONS AND RECOMMENDATIONS
The germanium doped fuse silica fibers, which were found to be the most
resistant to radiation induced transmission losses, were developed primarily for
their very low intrinsic attenuations so that they could be used as transmission
lines in lengths of the order of kilometers. These fibers appear at present to be
the most suitable for use where long lengths are required such as in ground based
communication and data transmission systems. They have some disadvantages,
however. Their cost per unit length is about ten times that of other glass fibers;
they are less flexible than other types of fiber, which causes a higher breakage
rate; and their numerical aperture is lower, making efficient coupling to light
sources and detectors more difficult. These factors, particularly the higher
cost, make consideration of other types of fiber worthwhile in applications, such
as in aircraft control and communication systems, where continuous lengths of
only a few tens of meters are required.
The plastic fibers, which showed a higher initial rate of radiation induced
transmission losses than the germanium-doped fused silica fibers, were found to
be superior to the others in their rate of recovery from transient x-ray induced
losses and their1resistance to neutron induced losses. Their cost per unit length
is only about 10 percent of that of the lowest priced glass fibers. These factors
may make them useful in systems in which their less desirable properties can be
tolerated. One of these properties is their extremely high attenuation in the infra
red, which makes them practically useless in this region of the spectrum where
operation is desirable because of the availability of more efficient solid state
light sources and detectors and because of greater security. The plastic fibers
have poor thermal properties, their upper temperature tolerance limit being only
about BOoC. They have low resistance to chemicals and moisture, and their
intrinsic losses at their optimum operating wavelengths are high (1200-1500 dB/km).
The cerium doped glass fibers show some potential for application in manned
systems. The 0.3 percent cerium doped fiber tested showed an induced change in
transmission of -125 dB/km (a decrease to 42 percent of the original transmission
for a 30 m length) after receiving an x-ray dose of 2000 ,rads, the assumed tolerance
level for manned systems. These fibers have all of the cost advantages and favor
able physical properties of standard commercial glass fibers, but they have
relatively high intrinsic attenuations of over 1200 dB/km.
The significance of the fluorescent pulses generated in the fibers during the
transient x-ray effects tests should be considered in view of the fact that, as
mentioned in paragraph 2.2 "Experimental Setup", the photodetector was shielded
by 10 cm of lead during the tests. This was necessary because even though the
detector was located in an area where it received doses from the x-ray pulses
53
more than two orders of magnitude lower than the fibers that were being tested,
without the shielding the radiation produced intense signals from the photodetector
similar to those observed during optical overloading. Since it is unlikely that
phtodetectors used in systems employing fiber optics would be so heavily shielded,
the magnitudes of false signals caused by radiation induced fluorescence in the
fibers may be of little consequence compared to the false signals produced in the
detectors.
Considering the above, the following recommendations are made:
(1) Continue development of the germanium doped fused silica fibers to lower
their cost and improve their mechanical properties.
(2) Develop cerium doped glass fibers with intrinsic attenuations below
1000 dB/km.
(3) Investigate new materials that could be used to produce plastic fibers
with improved optical and thermal properties.
(4) Develop photodetectors which maximize the ratio of optical sensitivity
to x-ray sensitivity.
These conclusions and recommendations are based only on the limited tests
reported here. There are many other aspects of the problem of the effects of
radiation on fiber optics, and optical materials in general, that mus t continue to
be investigated in order to improve our understanding of the interaction of nuclear
and space radiation with fiber optic systems.
54